Research And Application Of Ensemble Clustering Based On P System

The traditional clustering method adopts a single clustering algorithm that can only target a specific data set.If the true distribution of the data set does not meet the pre-specified hypothesis,the clustering result often fails to reflect the true distribution of the data set.Faced with heterogeneous datasets of various shapes and structures,ensemble clustering can achieve better average performance by integrating multiple base clustering results,which effectively avoiding single clustering algorithm results from noise,isolated points,and the sensitivity of sampling changes,etc.,and can search a new clustering result from multiple base clusters.This result is superior to any base clustering result.Membrane computing is a branch of the field of biological computing.The membrane computing model is also called the membrane system.The membrane system is a distributed and highly parallel computing system.Many studies have shown that in theory,many simple membrane systems have comparing with the Turing machine's computing power,and possibly exceeding the Turing machine in the future,the studies of membrane computing have become a hot research direction in bio-computing.In this paper,ensemble clustering optimization and membrane computing are studied.(1)This paper improves the K-means-based clustering integration algorithm.For the existing K-means-based clustering integration algorithm(KCC)to treat all cluster members fairly when performing cluster integration.Members of the base clusters treated fairly and did not take into account the different contributions of different members to the results.This paper designed an integrated member weight measurement method based on the mutual information theory(NMI)and designed it based on the degree of mutual information between cluster members and all other members.The weights are presented and the WIKCC algorithm is proposed.Experiments show that the improved algorithm has higher clustering accuracy.(2)This paper improves the clustering algorithm based on genetic algorithm(CEGA).According to the encoding method of CEGA algorithm based classification tags,this paper proposes a coding method based on micro-clusters.Every data point in a class member that is divided into the same cluster is treated as a micro cluster,that is,when the same data sample is processed to reduce the probability that they are separated during mutation and crossover,to improve the accuracy of the algorithm,and then use the classification label to perform chromosome coding on the micro-cluster.Finally,the membrane structure and membrane rules were designed to implement an improved algorithm.A GA-based Membrane Evolutionary Algorithm(GMEAEC)was proposed for cluster integration.Finally,it was experimentally shown to be improved.The clustering quality of the algorithm is improved,and the integration with different base clusters verifies that the algorithm is more robust than other comparison algorithms.(3)The WIKCC algorithm proposed in this paper is applied to image segmentation.The color features of the image are used as attribute features.The purpose of image segmentation is to extract useful information from the image.This paper measures the effectiveness of the algorithm by separating the degree of the object from the background.Comparing the WIKCC algorithm with KCC and K-means algorithm,from the visual effect of segmentation,the experimental results show that the segmentation result of WIKCC algorithm is obviously better than the other two algorithms.(4)This paper uses the GMEAEC algorithm to cluster oral disease data in the intelligent diagnosis and diagnosis,uses the best word segmentation tool currently applied to Python to segment the text data set,and uses TF-IDF to vectorize the data,and finally uses the TF-IDF to accurately digitize the data.The accuracy rate,recall rate,and F1-measure were used to measure the overall effect of clustering.The clustering efficiency was measured by the time the algorithm was run.Experiments show that the algorithm proposed in this paper is superior in clustering quality and efficiency in the clustering on oral diseases data than other algorithms.